By Sara Bronwen Hunter, Julie Weeds and Fiona Mathews
Data on human-nature interactions can be collected from online sources, such as social-media sites, but unlike data collected in field-based academic studies, they are extremely varied in content and fraught with potential biases. This can make understanding where these interactions (both positive and negative) are occurring in space more challenging. This prevents us from effectively directing resources to locations where conservation action is most needed.
In this paper, we explored whether using statistical methods from species distribution modelling could be used to predict the occurrence of human-nature interactions. We term this approach ‘human-nature interface mapping’. To evaluate the effectiveness of our approach, we used the case study of bat (Chiroptera) hunting, trade and consumption in Indonesia. We collected 475 records of these interactions from social media and online news using searches in Indonesian and English.
Overall, we were able to predict with high accuracy the occurrence of bat hunting, trade and consumption across Indonesia. The models developed in our study indicated that this potential threat to bat populations is highly associated with human populations and settlements, and that hunting occurs in different locations to trade and consumption. Our results highlight that even when using messy data from online sources, we can predict the spatial occurrence of human-nature interactions. The predictions from our study can be combined with locally relevant information to directly inform bat conservation in Indonesia.
Relational Thinking 